@omarvvvr provenance is quite important.. If you don't know what the function does, do you trust it?
especially in SME & corporate you should be able to track which lines were written by humans and by LLMs.
check this opensource project out: https://t.co/zo8tc6lT80
FSD NL: 3.5x fewer collisions is the headline. 1.6x is the non-highway number, where most serious injuries happen. Opt-in drivers skew cautious. Tesla is the only auditor. Number is real. Framing is generous.
This happens when unqualified people make decisions.. same level as putting in a single prompt in GLM5.2 to create a parody video and posting it without revisions.
BREAKING: The Administration has banned Anthropic's Fable after officials determined it found "a general cheat code capable of hacking any system."
The code has been recovered:
▲ ▲ ▼ ▼ ◄ ► ◄ ► B A START
The model declined to comment.
@steipete Makes sense to automate it that way.. but not a very efficient use of tokens for regular users..
In case you want to share your API key(s), I'm volunteering to accept them :)
@_ARahim_@steipete@bcherny What's the problem you're solving....? you're adding additional tokens without a need for it. the models understand it perfectly fine with typo's. Get your basic education in LLMs right, youtube it
@AnnickDeRidder@Tesla@AnnickDeRidder Dank! Het werd bijna irritant om telkens bij de grensovergang (NL->BE) het stuur over te nemen.. op naar de volgende stap: volledig autonoom
@thsottiaux I can't be the only one being frustrated to have close to zero visibility when and how my extra codex usage credits are being spend. Please fix this with decent settings